© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 41926
    Prev
    Next
    INDEX
    Explanations

    actions related to removal or loss

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    iffies
    -0.15
    .scalablytyped
    -0.15
    رخ
    -0.14
    someone
    -0.14
    /how
    -0.14
     alguien
    -0.14
     Leban
    -0.13
    /from
    -0.13
    ัà¸ļà¸ľ
    -0.13
    گاÙĩÛĮ
    -0.13
    POSITIVE LOGITS
     all
    0.50
     any
    0.43
     most
    0.39
    æīĢæľī
    0.33
     everything
    0.33
    ãģĻãģ¹ãģ¦
    0.32
     semua
    0.32
    ä¸ĢåĪĩ
    0.32
     ALL
    0.31
     altogether
    0.30
    Activations Density 0.198%

    No Known Activations